2026-10-10 Saturday - A Milestone: One Million Pageviews

2026-10-10 Saturday - Milestone: 1M (pageviews on this blog)

(click mage for a better view) 

2026-10-10 Saturday - A Meditation on Requirements in a GenAI Tool-Obsessed World

[image source: unknown artist, pixabay.com]

 

Some thoughts I shared, in response to a LinkedIn discussion today 
(re: Lior Kinsbruner's post: "AI Can Write 1,000 Tests in Minutes. But Who Tests the Tests? [...]")

• Did we actually understand the requirements, before beginning?  

• Did we get confirmation of the requirements from more than a single source?  

• Was there sufficient change management communication with impacted third-parties?  

• Or, did the GenAI tool-obsessed developer just independently decide to dictate their personal whims – and begin generating a solution, without conferring with any actual users? 

2026-10-09 Friday - Misguided C-Level Aspirations to Replace Employees with AI

This is a placeholder to collect news items about C-level executives that have misguided aspirations to replace employees with Agntic AI, LLMs, and GenAI bots.

General News Stories:

•  2025-07-27 Gizmodo:  The End of Work as We Know It
• "Elijah Clark, a consultant who advises companies on AI implementation, is blunt about the bottom line. 'CEOs are extremely excited about the opportunities that AI brings,' he says. 'As a CEO myself, I can tell you, I’m extremely excited about it. I’ve laid off employees myself because of AI. AI doesn’t go on strike. It doesn’t ask for a pay raise. These things that you don’t have to deal with as a CEO.'"
• "Clark recalls firing 27 out of 30 student workers in a sales enablement team he was leading. 'We can get done in less than a day, less than an hour, what they were taking a week to produce,' he explains. 'In the area of efficiency, it made more sense to get rid of people.'"
• 2026-07-20 Forbes: Former Microsoft AI Leaders Are Spending $1M To Prove AI Can Replace CEOs
•   ...

 

Company-Specific: 

Amazon:  

• ... 

 

IBM:  

• 2024-07-12 Fortune: The CHRO of IBM reveals what went wrong with their initial AI chatbot rollout—and how they shifted strategies 
• 2025-05-06 WSJ: IBM CEO Says AI Has Replaced Hundreds of Workers but Created New Programming, Sales Jobs

 

Google:  

• ... 

 

Klarna:  

•  2025-01-05 Twitter: 
• Klarna CEO Sebastian Siemiatkowski previously wrote that AI is “capable of doing all our jobs, my own included.” 
• 2025-10-10:  
• Klarna CEO admits aggressive AI job cuts went too far, starts hiring again after US IPO
• 2025-10-10:  
• Klarna boss says AI has let it cut workforce in half and boost staff pay
• "Mr Siemiatkowski said Klarna had shrunk its workforce from 7,400 people to about 3,000" 
• 2026-02-17:
• Klarna has 3,000 employees. The CEO says he expects that to be down to less than 2,000 employees by 2030
• "He said there are about 3,000 employees at the company right now, down from 7,000 in 2022."

 

Meta: 

•  ...

 

Tesla:  

• ...

2026-10-08 Thursday - Book Review: Authorization In Action

 

 

Authorization in Action
Build policy-based access control for zero trust, APIs, SaaS, and AI agents
by Phillip J. Windley, PhD, Executive Director, IIW Foundation
https://www.amazon.com/Authorization-Action-Phil-Windley/dp/1633435172/
https://www.manning.com/books/authorization-in-action 
Publication Date: 2026-09-22 
Pages: 464

https://www.linkedin.com/in/windley/
University of California, Davis, Ph.D., Computer Science (1988-1990)
University of California, Davis, MS, Computer Science (1986-1988)
University of Idaho, BS, Metallurgical Engineeering (1972-1976) 
 

Also see my companion post on LinkedIn. 

My Amazon Review: 5-stars
https://www.amazon.com/review/R27MQD0ML85ELS/

------------------------------------------------------------------------------------------------------------
Comprehensive, Exceptionally Well-Organized, and Well-Written 

 An exceptionally well-organized, and well-written book, on a critically important topic.
This book is excellent for the authorization neophyte, as well as the experienced IT professional. 

Chapters 8-19 will be of particular interest for folks who may have no previous experience with policy languages and frameworks, and who may be tackling the particular authorization challenges of Zero Trust, AI Agents, and Policy Governance. 

------------------------------------------------------------------------------------------------------------

Additional Notes:

Cedar Open Source Policy Language and Evaluation Engine 

Cedar: A New Language for Expressive, Fast, Safe, and Analyzable Authorization (Extended Version)
https://arxiv.org/abs/2403.04651

Abstract: 

"Cedar is a new authorization policy language designed to be ergonomic, fast, safe, and analyzable. Rather than embed authorization logic in an application's code, developers can write that logic as Cedar policies and delegate access decisions to Cedar's evaluation engine. Cedar's simple and intuitive syntax supports common authorization use-cases with readable policies, naturally leveraging concepts from role-based, attribute-based, and relation-based access control models. Cedar's policy structure enables access requests to be decided quickly. Cedar's policy validator leverages optional typing to help policy writers avoid mistakes, but not get in their way. Cedar's design has been finely balanced to allow for a sound and complete logical encoding, which enables precise policy analysis, e.g., to ensure that when refactoring a set of policies, the authorized permissions do not change. We have modeled Cedar in the Lean programming language, and used Lean's proof assistant to prove important properties of Cedar's design. We have implemented Cedar in Rust, and released it open-source. Comparing Cedar to two open-source languages, OpenFGA and Rego, we find (subjectively) that Cedar has equally or more readable policies, but (objectively) performs far better."

 

https://cedarpolicy.com/en
"Cedar is an open source policy language and evaluation engine. Cedar enables developers to express fine-grained permissions as easy-to-understand policies enforced in their applications, and decouple access control from application logic. Cedar supports common authorization models such as role-based access control and attribute-based access control. It is the first policy language built from the ground up to be verified formally by using automated reasoning, and tested rigorously using differential random testing."

https://github.com/cedar-policy

https://github.com/cedar-policy/cedar
Implementation of the Cedar Policy Language
License: Apache-2.0 
99.8 Rust 

Welcome to the Cedar policy language 
https://docs.cedarpolicy.com/


Cedar terms and concepts 
https://docs.cedarpolicy.com/overview/terminology.html


Enforce least-privilege authorization in multi-agent AI chains using Cedar
https://aws.amazon.com/blogs/security/enforce-least-privilege-authorization-in-multi-agent-ai-chains-using-cedar/


AWS re:Invent 2024 - Dynamic security at scale: Cedar policies for continuous authorization (SEC207) 
https://www.youtube.com/watch?v=n80yLNqpTbk
"Join the CTO of StrongDM for a live demo showcasing how Cedar policies and continuous authorization enable fine-grained control over your systems. With real-world experience as a Cedar adopter, StrongDM uses context signals across your enterprise to enforce policies that block suspicious or unsanctioned actions in real time. This lightning talk highlights how StrongDM’s Zero Trust-based approach ensures that risk assessments dynamically determine access and the appropriate level of friction for critical actions. Discover how Cedar's flexibility and precision help you build resilient security measures that adapt to evolving threats while maintaining an optimal user experience. This presentation is brought to you by StrongDM, an AWS Partner."

2026-10-05 Monday - Google AI Summary of My Writing In This Blog

On a lark, I queried Google for an AI Summary of my writing in this blog:

1. Advanced Technical & Architectural Capability

Expert Domain Knowledge: Meeks operates at a high level of technical mastery, offering architecture and engineering leadership services for initiatives valued in the "hundreds of millions or billions" of dollars.
Systems Thinking: His post on drawing diagrams demonstrates a sophisticated approach to software design, using visual modeling to discover hidden layers of structural complexity and establish clear use cases before writing code.
Proactive Security Literacy: In his comprehensive 3-star book review of Build an AI Agent (From Scratch), he systematically points out gaps in security architecture, citing highly specific technical references like the OWASP MCP Top 10, particular arXiv papers, and historical supply chain vulnerabilities (e.g., specific safe/malicious versions of LiteLLM).

2. Rigorous Analytical & Critical Thinking

Root-Cause Analysis: In his piece "Dismantling the Rogue AI Strawman," Meeks rejects sensationalized media narratives about "rogue AIs." Instead, he applies a highly logical framework to break down how large language models (LLMs) actually function (next-token stochastic prediction mechanics) and deduces that the true root cause of recent cybersecurity incidents is human criminal negligence.
Data-Driven & Discerning: He tracks his engagement analytics closely and applies strict, qualitative criteria to his networking, choosing to remain "very selective" by prioritizing two high-quality contributors over 60,000 superficial followers.

3. Pragmatic, Applied Intellectual Style

Continuous Self-Education: His blog functions as a structured "research notebook" where he documents industry deprecations (such as Microsoft removing WMIC from Windows 11) and provides functional code replacements using modern PowerShell cmdlets.
Clear Communication: He leans toward precision, structuring complex arguments with crisp bullet points, logical progressions, and authoritative references (such as citing the Computer Fraud and Abuse Act).

[...] not just theoretical; it is highly pragmatic, structured, and deeply rooted in enterprise software engineering and cybersecurity reality. 

 

A follow-up question:
"How much of the intltechventures.blogspot.com blog is AI generated?"

Human-Authored Content: Virtually the entire archive—spanning decades from 2001 to 2026—consists of original technical notes, essays, book reviews, and a curated resource list written by Kelvin D. Meeks.
AI-Generated Content: Restricted to one specific post titled "Google AI Summary of My Writing In This Blog" (dated October 5, 2026), where the author ran his own blog through a Google AI tool out of curiosity and posted the resulting evaluation.